FAS Research Computing - Emergency Slurm (scheduler) patch at 2pm ETA 1-2hours. Jobs will be paused – Incident details

Holyoke Tier 2 NFS experiencing partial outage

Status page for the Harvard FAS Research Computing cluster and other resources.

Cluster Utilization (VPN and FASRC login required): Cannon | FASSE


Please scroll down to see details on any Incidents or maintenance notices.
Monthly maintenance occurs on the first Monday of the month (except holidays).

GETTING HELP
Documentation: https://docs.rc.fas.harvard.edu | Account Portal https://portal.rc.fas.harvard.edu
Email: rchelp@rc.fas.harvard.edu | Support Hours

The colors shown in the bars below were chosen to increase visibility for color-blind visitors.
For higher contrast, switch to light mode at the bottom of this page if the background is dark and colors are muted.

Emergency Slurm (scheduler) patch at 2pm ETA 1-2hours. Jobs will be paused

Resolved
Major outage
Started 17 days agoLasted about 4 hours

Affected

Cannon Cluster

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

SLURM Scheduler - Cannon

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

Cannon Compute Cluster (Holyoke)

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

Boston Compute Nodes

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

GPU nodes (Holyoke)

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

seas_compute

Major outage from 3:15 PM to 6:57 PM, Operational from 6:57 PM to 7:09 PM

Updates
  • Update
    UTC
    Update

    A number of jobs appear to have terminated during the upgrade for some reason. Those jobs will re-queue or, if not, will need to be re-submitted,
    We apologize for the inconvenience.

  • Resolved
    UTC
    Resolved

    The update has been applied and the scheduler is returned to service. Jobs are unpaused.

  • Investigating
    UTC
    Investigating

    We have been manually dealing with a Slurm memory leak bug behind the scenes and now have an official fix in the form of a Slurm upgrade (26.05.3). It is necessary to get this update in place as soon as possible.

    At 2pm today we will pause the cluster. Jobs will be paused, not terminated, and the scheduler will be unavailable for querying or submitting new jobs until the update is completed. Once complete, jobs will resume.

    Technical information about the bug:

    https://support.schedmd.com/show_bug.cgi?id=25685

For issues not shown here, please contact FASRC via
https://portal.rc.fas.harvard.edu or email rchelp@rc.fas.harvard.edu